Blur Invariant Phase Correlation in X-Ray Digital Subtraction Angiography

Digital subtraction angiography (DSA) is a very useful tool to examine and visualize coronary blood vessels, especially for the coronary interventions. However coronary X-ray digital subtraction angiography suffers from heart beat and patient motion artifacts, and current commercial DSA devices don´t provide subtraction images for the coronary DSA since the currently algorithms cannot be fast enough. In this paper, a new approach to the registration of coronary X-ray digital subtraction angiographic image sequences is proposed. The approach combines blur invariant phase correlation and thin-plate spline image warping by means of template matching technique. The proposed algorithm is fully automatic and very efficient to remove the motion artifacts. It has been applied to the simulated data and several clinical data sets. The experiment results show that the proposed algorithm is robust and fast.